1 00:00:00,060 --> 00:00:02,610 We want to thank you both for coming. 2 00:00:02,910 --> 00:00:09,260 I want to thank the guys and the team that advertised this week that I think will ensure that our advertising and it's 3 00:00:09,500 --> 00:00:15,959 so it's great to see you already to extend the numbers and get 45 bookings and then rolling out to have extended to 100. 4 00:00:15,960 --> 00:00:20,430 So I'm sure that's for the whole of the of the of the speakers that we've got tonight, which I'm very, 5 00:00:20,430 --> 00:00:27,720 very pleased to be able to to announce that we've got three great speakers who all, I'm sure will introduce themselves as they go along. 6 00:00:28,770 --> 00:00:31,770 On the topic of same sex ed and what's the book about them? 7 00:00:32,040 --> 00:00:35,580 The Ghost of Christmas, Christmas, Present, Christmas, The Past and the Future. 8 00:00:35,940 --> 00:00:38,880 So you look under the Christmas carols reviews tonight. 9 00:00:39,960 --> 00:00:44,010 Without further ado, I will I will introduce Ian Chalmers is going to start us off on the clock. 10 00:00:46,650 --> 00:00:50,219 Well, I'm not Scrooge least I hope I'm not Scrooge. 11 00:00:50,220 --> 00:00:55,920 I never thought of myself that way. But I'm going to talk about the past quite a long way into the past. 12 00:00:55,920 --> 00:00:59,550 In fact, as you see from the title of the first slide, 13 00:01:00,870 --> 00:01:07,980 I'm going to cover extremely fast from the beginning of the 18th century, the beginning of the 21st century. 14 00:01:08,670 --> 00:01:12,420 So let me just start off with a sort of personal declaration. 15 00:01:15,810 --> 00:01:22,800 It is. Why is evidence synthesis to assess treatment effects so important? 16 00:01:22,910 --> 00:01:27,650 It's because without it, people suffer and die unnecessarily. 17 00:01:28,820 --> 00:01:37,910 Quite a long time ago, an article that I've written was published in the BMJ asking the questions What do I want from health research and researchers? 18 00:01:37,910 --> 00:01:47,450 When I'm a patient and obviously I want the evidence that's taken into account when I'm being treated to to be relevant and reliable. 19 00:01:47,990 --> 00:01:54,110 And for me, that will usually mean evidence from systematic reviews of research designed 20 00:01:54,110 --> 00:02:00,530 to minimise the effects of biases of various kinds and of the play of chance. 21 00:02:04,120 --> 00:02:04,780 In fact, 22 00:02:05,170 --> 00:02:14,590 measures to reduce the effects of bias in the play of chance when assessing the effects of treatment has been a concern since at least since 1722. 23 00:02:15,730 --> 00:02:22,540 None of you in this room will have seen this disease. Does anyone not know what this disease is? 24 00:02:26,980 --> 00:02:31,060 Okay, David, this is smallpox, and none of you will have seen it. 25 00:02:31,090 --> 00:02:33,580 It's a horrible condition, has a high mortality. 26 00:02:35,890 --> 00:02:46,150 So in essence, the general practitioner in Halifax, a man called Nettleton, wrote to the secretary of the Royal Society, 27 00:02:47,140 --> 00:02:54,880 James Durant, giving some data that they had on the mortality associated with naturally acquired smallpox. 28 00:02:56,440 --> 00:03:09,940 And he because there was dispute about whether inoculation that's actually giving smallpox scabs as the inoculum was being talked about in those days. 29 00:03:11,980 --> 00:03:21,940 He said, what we need is a comparison between the mortality associated with the inoculation and the mortality associated with the natural disease. 30 00:03:23,500 --> 00:03:28,270 And he collected data locally in Yorkshire. 31 00:03:29,560 --> 00:03:36,970 And what you'll see already is that that bottom line, the total, he's actually, in essence, done a crude meta analysis. 32 00:03:36,980 --> 00:03:49,000 He's taken these different studies from different areas and found an average effect describing the the totality. 33 00:03:52,380 --> 00:04:04,920 What he also said was that not a single one of the 60 or so I think it was, who had had inoculation, had died of smallpox. 34 00:04:05,670 --> 00:04:16,170 And no one needs any statistical expertise to know that that's likely to be it's highly unlikely to be a chance difference. 35 00:04:18,760 --> 00:04:25,930 So he, for a start recognised the needs for comparisons and sufficient numbers of observations. 36 00:04:27,640 --> 00:04:35,470 However, Isaac Masi was someone who I think I gather is it's mainly for religious reasons, 37 00:04:35,480 --> 00:04:42,910 didn't like the idea of man intruding in God's plans for people. 38 00:04:44,110 --> 00:04:53,240 But he made a good point, he said. What you've got to do is to make sure it like is being compared with like in any comparison. 39 00:04:54,930 --> 00:05:03,540 To form a just comparison and calculate, right? In this case, the circumstances of the patients must and ought to be as near as may be on a par. 40 00:05:06,000 --> 00:05:09,000 So he recognised the needs for fair comparisons. 41 00:05:10,520 --> 00:05:17,540 James Durin, Secretary of the Royal Society, recognised the need to keep systematic reviews up to date. 42 00:05:19,810 --> 00:05:28,360 So he and 1724 published an update of net Nettleton Yorkshire data, 43 00:05:29,050 --> 00:05:37,690 but added to IT data that he had received from Chichester and Haverfordwest and the in. 44 00:05:37,690 --> 00:05:43,090 I think there were 80 cases who had been inoculated up from 60. 45 00:05:44,260 --> 00:05:52,450 And there again, there was not one case of smallpox associated with the inoculated group. 46 00:05:55,120 --> 00:06:05,110 Now, a couple of years later, another fellow of the Royal Society, his ex, certainly His Excellency, custom ATA ambassador from Tripoli. 47 00:06:06,760 --> 00:06:09,790 He produced some evidence from North Africa. 48 00:06:11,410 --> 00:06:16,390 And what he said was out of 100 persons inoculated not to die. 49 00:06:16,420 --> 00:06:16,899 In other words, 50 00:06:16,900 --> 00:06:26,800 not not even as much as two people die out of 100 persons that are infected with the smallpox in the natural way that I commonly about 30. 51 00:06:27,550 --> 00:06:35,020 So even higher mortality rate than had been shown in the British examples. 52 00:06:36,540 --> 00:06:46,980 And he pointed out that inoculation is generally practised not only by the inhabitants of the times, but also by the Wild Arabs. 53 00:06:48,030 --> 00:06:54,030 Now, how one takes that into account in an updated systematic review, one doesn't quite know, 54 00:06:54,030 --> 00:07:00,750 but it certainly would want to be mentioned because it refers to a different part of the world and that might be relevant to know that. 55 00:07:02,670 --> 00:07:07,140 So. Choice. 56 00:07:07,310 --> 00:07:17,030 John Gasp A chaser actually updated James during this analysis after James Jones had died again. 57 00:07:17,030 --> 00:07:31,010 1729 And you see the rather nice table he's produced there both for the this is for between 1721 and 1728, both for associated with natural smallpox. 58 00:07:31,430 --> 00:07:42,590 And he shows you the proportion there, the poorer proportion, his calculated and then the those that have it by inoculation. 59 00:07:44,990 --> 00:07:56,090 So this document issued June last year, one can only say about it is better 289 years late than never. 60 00:07:57,410 --> 00:08:04,640 It's not just the Royal Society that, as far as I know, has not shown much of an interest in this issue at all. 61 00:08:04,970 --> 00:08:11,090 Certainly, I'm confident in saying that the Academy of Medical Sciences has not only not shown any interest, 62 00:08:11,120 --> 00:08:16,490 it has actually blocked attempts to generate interest within that organisation. 63 00:08:19,190 --> 00:08:21,850 You would have thought it was a rather straightforward issue. 64 00:08:21,860 --> 00:08:32,990 But clearly some people have felt challenged, let's put it that way, by the development of methodology for this type of research. 65 00:08:35,710 --> 00:08:40,810 What about pioneering the development and reporting of methods used for research syntheses? 66 00:08:41,980 --> 00:08:46,990 Well, here we can bring in this Scottish navel surgeon, James Lind, 67 00:08:47,470 --> 00:08:53,890 who's usually known for the controlled trial that he did aboard a ship called HMS Salisbury, 68 00:08:54,940 --> 00:09:02,979 comparing six different ways of treating scurvy and demonstrating really beyond any reasonable doubt that oranges, 69 00:09:02,980 --> 00:09:08,110 lemons were a rather good idea compared with the other things that were being used. 70 00:09:08,110 --> 00:09:14,050 For example, the College of Physicians favourite approach was sulphuric acid. 71 00:09:16,080 --> 00:09:26,530 So there were a number of steps that he took to try and do a decent job of reviewing what was known. 72 00:09:27,790 --> 00:09:32,950 And you'll see that on the title page of his book, 73 00:09:33,430 --> 00:09:39,940 he mentions that it includes a critical and chronological view of what has been published on the subject. 74 00:09:41,680 --> 00:09:48,220 He went on in the introduction to say, as it is, no easy matter to root out prejudices, biases. 75 00:09:48,610 --> 00:09:54,100 It became requisite to exhibit a full and impartial view of what had hitherto been published on the scurvy. 76 00:09:54,610 --> 00:10:00,459 And that in a chronological order by which the sources of these mistakes may be detected. 77 00:10:00,460 --> 00:10:06,940 Indeed, before the subject could be set in the clear and proper light, it was necessary to remove a great deal of rubbish. 78 00:10:08,080 --> 00:10:11,620 And this gives some idea of the care that he took to do this. 79 00:10:12,070 --> 00:10:14,860 Now I've got a copy of a facsimile, in fact, 80 00:10:16,540 --> 00:10:30,100 here of his book and mark the page where he describes the efforts that he went to to obtain the relevant information. 81 00:10:30,400 --> 00:10:33,160 So you might like to pass that around during the course of the evening. 82 00:10:37,760 --> 00:10:45,290 Now I'm going to jump forward quite a long way but hundred and 50 years because it was really. 83 00:10:48,530 --> 00:10:59,150 A physicist. Who made it clear that scientists should accumulate scientifically because science is cumulative. 84 00:10:59,480 --> 00:11:06,020 Sometimes there are big breakthroughs, in fact, where you ditch everything that has been thought of as important done then. 85 00:11:06,020 --> 00:11:09,920 But more usually it's a cumulative process. 86 00:11:12,130 --> 00:11:14,950 John Strutt. Lord Reilly was a physicist. 87 00:11:14,970 --> 00:11:28,330 He discovered Argon and he was president of the Association for the Advancement of Science, and they had a two month meeting in Montreal in 1884. 88 00:11:29,140 --> 00:11:40,240 Let's get proper attention to things. If you have met meetings that last couple of months and he said in his presidential address, if, 89 00:11:40,240 --> 00:11:47,020 as is sometimes supposed science consisted in nothing but the laborious accumulation of facts, it would soon come to a standstill. 90 00:11:47,260 --> 00:11:49,270 Crushed as it were, under its own weight. 91 00:11:50,140 --> 00:11:58,240 The work which deserves, but I'm afraid, does not always receive the most credit is that in which discovery and explanation go hand in hand, 92 00:11:58,600 --> 00:12:04,300 in which not only are new facts presented, but their relation to old ones is pointed out. 93 00:12:05,970 --> 00:12:13,800 Austin, Bradford Hill. Probably Britain's greatest medical statistician approached it rather differently. 94 00:12:13,810 --> 00:12:16,980 He said, What do readers want to know in the report of a study? 95 00:12:17,580 --> 00:12:23,910 They want to know why. Why was it started? What did the authors do? 96 00:12:23,940 --> 00:12:26,370 What answer did they get? And what does it mean? 97 00:12:27,510 --> 00:12:35,700 And that implies that reports of new research should begin and end with systematic reviews of what is already known, 98 00:12:36,540 --> 00:12:44,370 so that results are set in context. And failure to do it results in avoidable suffering and death. 99 00:12:46,580 --> 00:12:53,570 So going back to this issue, we've spoken I've spoken about some of the biases that can enter in. 100 00:12:54,230 --> 00:12:57,290 But what about reducing the player of chance? 101 00:12:58,640 --> 00:13:09,320 Well, this is an early example published, as you see in 1904, which takes a number of studies of the effects of inoculation against time, 102 00:13:09,860 --> 00:13:18,380 typhoid, enteric fever in the British army and calculates a mean value taken from the individual studies. 103 00:13:19,880 --> 00:13:25,010 This one a little bit later on shows in the. 104 00:13:27,620 --> 00:13:33,499 Data I outlined in red the cumulated results. 105 00:13:33,500 --> 00:13:43,130 The metre analysed results of treatment serum treatment for lobar pneumonia in three separate hospitals in New York, 106 00:13:43,430 --> 00:13:49,940 where alternation had been used to decide who was going to get the serum and who was not. 107 00:13:51,170 --> 00:13:58,520 And as you will see, statistical estimates of variance start to get mentioned by then. 108 00:13:58,640 --> 00:14:08,270 Toes. I've put Isabella Leach in there, not because she has, as far as I know, contributed to the ideas in that analysis. 109 00:14:08,270 --> 00:14:14,390 But it seemed to me it was about time that a great woman and indeed she was featured in my slides. 110 00:14:15,380 --> 00:14:18,260 This paper, which is available on the James Lynn Library, 111 00:14:18,500 --> 00:14:26,870 is a beautiful description of the rationale for and the steps needed to be taken to do a decent job in systematic reviews. 112 00:14:28,670 --> 00:14:37,700 Donald Reed used individual patient data from nine controlled trials to do a more powerful, 113 00:14:37,700 --> 00:14:44,960 statistically powerful assessment of the role of anticoagulants in myocardial infarction. 114 00:14:46,280 --> 00:14:49,670 Jan stands up in 1974. 115 00:14:51,110 --> 00:14:57,230 He looked at controlled trials of adjuvant radiotherapy in women who had had surgery for breast cancer. 116 00:14:59,230 --> 00:15:02,240 To the right of therapy. It was in general to the axilla. 117 00:15:03,410 --> 00:15:09,350 And what you showed was that in the controlled trials that were done, these were all randomised trials. 118 00:15:10,340 --> 00:15:16,880 There was a higher death rate in the women who had had the radiotherapy. 119 00:15:18,210 --> 00:15:27,060 And this effect was particularly evident in women who had surgery for their left breast cancer. 120 00:15:28,230 --> 00:15:31,380 Some of you may be thinking about how that could be. 121 00:15:32,580 --> 00:15:38,239 I won't mention it. Some of you may have guessed already. Archie Cochrane said. 122 00:15:38,240 --> 00:15:42,620 It's surely a great criticism of our profession that we haven't organised a critical summary. 123 00:15:42,750 --> 00:15:52,280 That was the term that he used for systematic review adapted periodical periodically kept up to date of all relevant randomised trials. 124 00:15:55,540 --> 00:16:06,040 The recognition came in the 1980s within medicine of the importance of global collaboration for evidence synthesis. 125 00:16:07,000 --> 00:16:13,390 So you get this example published in the New England Journal of Medicine on the effects of adjuvant, 126 00:16:13,750 --> 00:16:17,290 Tamoxifen and cytotoxic therapy for breast cancer. 127 00:16:18,150 --> 00:16:19,860 This one published in the BMJ? 128 00:16:19,860 --> 00:16:29,220 I think, yes, in the BMJ from a largely Oxford group on the effects of antiplatelet treatments with cardiovascular diseases. 129 00:16:30,650 --> 00:16:41,600 And then this monstrous two volume book attempted to do a systematic review of basically the whole field of obstetrics and midwifery. 130 00:16:42,440 --> 00:16:52,580 And perhaps what was new about this was all of the data were published electronically and have been updated ever since then, electronically. 131 00:16:53,900 --> 00:16:59,180 Archie Cochrane very kindly gave that a fair wind. 132 00:16:59,390 --> 00:17:05,810 He wrote in the forward that he thought it was good and that coming from him was a great compliment. 133 00:17:09,180 --> 00:17:18,180 That's if you like. Pilot study led to the Cochrane Collaboration and the founding chair of the Cochrane 134 00:17:18,180 --> 00:17:25,170 Collaboration was a predecessor of Karl's as director of the Centre for Evidence based Medicine. 135 00:17:25,200 --> 00:17:37,110 He was the first director of the centre, in fact, and one should not assume that Cochrane reviews were particularly important in what 136 00:17:37,740 --> 00:17:42,330 was being recognised by people who realised that this form of research was essential. 137 00:17:42,990 --> 00:17:50,819 You'll see, sort of coursing along the bottom of these graphs is the numbers, numbers of Cochrane reviews. 138 00:17:50,820 --> 00:17:58,560 Over time you'll see this growth in. There's an interest in these methods far beyond the Cochrane Collaboration. 139 00:18:00,030 --> 00:18:12,540 I just want to give three examples for stopping of nice examples of the way systematic reviews can be developed and used for other purposes. 140 00:18:13,830 --> 00:18:28,650 This one shows how, if you're trying to find out what might persuade teenagers to eat five fruits a day instead of Mars bars and whatever else, 141 00:18:29,430 --> 00:18:37,050 then you can look and do a systematic review of the randomised trials that have been done to try and alter eating habits. 142 00:18:37,380 --> 00:18:42,600 But you can also do a systematic review of the qualitative data that arises after 143 00:18:43,230 --> 00:18:47,610 out of asking children themselves what they think might persuade them to change. 144 00:18:48,270 --> 00:18:54,809 And you can look at these both together, and this is why I think this paper published in the BMJ was so beautiful because it actually 145 00:18:54,810 --> 00:19:00,090 showed data from both different types of data coming from slightly different perspectives, 146 00:19:00,090 --> 00:19:09,540 but addressing the same fundamental problem. I found that, in fact, you can use systematic reviews to detect plagiarism and fraud. 147 00:19:10,620 --> 00:19:22,200 And this Croatian obstetrician, as in Korea, was demonstrated beyond any shadow of a doubt to have plagiarised other people and to. 148 00:19:25,180 --> 00:19:29,170 Have been fraudulent as well. He has gone from strength to strength. 149 00:19:29,530 --> 00:19:39,070 I might point out he now runs an extremely well-heeled operation in Dubrovnik and has paid no price for this at all. 150 00:19:40,600 --> 00:19:48,310 This one is so important because it means it's a it's a really beautiful example of how you can address additional hypotheses. 151 00:19:50,510 --> 00:19:54,330 To those that the studies were originally designed to address. 152 00:19:54,350 --> 00:20:01,280 All of these people are basically cardiovascular disease epidemiologists and clinicians. 153 00:20:02,540 --> 00:20:08,720 They thought that they might look at the to see because there was some evidence suggesting that 154 00:20:09,590 --> 00:20:15,950 aspirin might actually have an impact on cancer incidence and cancer and cancer treatment. 155 00:20:16,790 --> 00:20:17,629 And lo and behold, 156 00:20:17,630 --> 00:20:27,040 they found that a daily aspirin source on massive numbers of daily aspirin reduced deaths due to several common cancers during and after the trials. 157 00:20:27,050 --> 00:20:30,470 And it was it was about a 20% reduction. Big deal. 158 00:20:31,830 --> 00:20:41,580 Then I'll end with two acknowledgements of the importance of the group based at the Centre for Evidence based Medicine. 159 00:20:43,950 --> 00:20:55,800 This one is very important because and it may well be that that Carl refers to this, but I regard it as historically very important. 160 00:20:56,400 --> 00:21:02,520 Indeed, it shows how one can address the worry of publication bias. 161 00:21:03,030 --> 00:21:09,569 And lastly, this, I think is one of the most important articles written on this topic by Kim. 162 00:21:09,570 --> 00:21:21,030 I'll be speaking in a moment. I think it's an extremely sensible injunction that all health researchers should 163 00:21:21,030 --> 00:21:26,010 begin their training by preparing at least one systematic reviews when for review. 164 00:21:26,400 --> 00:21:37,940 When I asked. My son, who works in education research to comment on a draft of this presentation, he said, is you can delete the word health. 165 00:21:38,150 --> 00:21:41,570 All researchers should begin that training. Thank you very much indeed. 166 00:21:47,040 --> 00:21:50,020 If it's all right, we'll take questions at the end. Yeah. 167 00:21:50,190 --> 00:21:57,240 And then someone move straight on to, uh, let's call that again, talking to us about the presence of system activities. 168 00:21:58,170 --> 00:22:09,400 Thank you. It's quite hard to define the president, isn't it? 169 00:22:11,500 --> 00:22:18,730 So in turning the President, what I thought there is I just for how to go about it in mentioned a bit about the definition. 170 00:22:19,620 --> 00:22:23,940 And I've been thinking about the definition quite a lot recently. 171 00:22:23,950 --> 00:22:30,040 So I want you to keep in mind some things while I'm talking, the idea of a critical summer. 172 00:22:30,700 --> 00:22:36,220 What do we mean by that? What did she mean by that? By speciality or subspecialty? 173 00:22:37,030 --> 00:22:38,740 That's quite an interesting issue as well. 174 00:22:38,750 --> 00:22:47,260 Now, in a world of ageing and multimorbidity adapted periodically, that's quite hard to do and all relevant randomised controlled trials. 175 00:22:47,260 --> 00:22:51,729 So I'm in the present, I've been thinking a lot about that, so I thought I'd start about my journey. 176 00:22:51,730 --> 00:22:57,340 So I'm going back is the present. I'm in 15 years, so I'm only going to jump 15 years. 177 00:22:57,340 --> 00:23:03,069 Where is the jump to 150 years? And this is the first review I participated in. 178 00:23:03,070 --> 00:23:09,940 So I started this journey about 2003 and published this with Paul Glazier, who many will know. 179 00:23:09,940 --> 00:23:13,360 He's also a former director of the centre and Rafael PEREIRA. 180 00:23:14,140 --> 00:23:21,250 Interesting to go back 15 years. I always thought, you know, it's interesting when you think about evidence and what's happening, 181 00:23:21,250 --> 00:23:25,300 you think, well, surely this has been there for always and it's been there. 182 00:23:25,570 --> 00:23:31,240 But actually this is at the period where a lot of the formulation of the ideas were coming in the first ten years. 183 00:23:31,240 --> 00:23:36,459 Is that fair enough to say and look, the methods, that's what we did with no, no. 184 00:23:36,460 --> 00:23:42,220 Right now, if you do a Cochrane review, you pretty much then said, these are my methods, here's my protocol. 185 00:23:42,220 --> 00:23:50,530 And people said, get on with it. And people were using scales and the methods we use because that's how we approach the quality of trials. 186 00:23:51,550 --> 00:23:54,820 And we sort of did it low, moderate and high risk. 187 00:23:57,190 --> 00:23:59,860 That's pretty much what great looks like, isn't it, though, 188 00:23:59,860 --> 00:24:09,100 when you think about grade and then bring in this on top structure actually was pretty well thought through and set out in the early days. 189 00:24:09,370 --> 00:24:14,260 And I know I'm not sure if this helps you or hinders you, 190 00:24:14,710 --> 00:24:23,230 but actually some of the structures are a bit more if you're not like prescriptive were easier to get your review through than this sort of approach. 191 00:24:25,000 --> 00:24:32,170 Second as well, the basics of matter analysis of were and the systematic reviews were very clearly set out very early 192 00:24:32,170 --> 00:24:39,220 on and they haven't changed very much within your single traditional review of an intervention, 193 00:24:39,880 --> 00:24:46,570 a a statistical mechanism combining data with the forest plot, a type of effect model, 194 00:24:47,080 --> 00:24:53,200 fixed or random heterogeneity assessment funnel plots, and then the interpretation with caution. 195 00:24:56,040 --> 00:25:00,360 So that's what I'm calling the first generation of my review journey. 196 00:25:00,810 --> 00:25:06,240 This is then the second generation of my review journey, which is expanding, realising that. 197 00:25:06,930 --> 00:25:11,970 And we published something in Cochrane and this is about self-monitoring of oral anticoagulation. 198 00:25:12,330 --> 00:25:16,350 But actually at the same time you could get things out there earlier in a journal. 199 00:25:16,650 --> 00:25:21,240 And this is the same publication actually about a year early in The Lancet. 200 00:25:22,590 --> 00:25:32,850 At the same time, in fact, actually previous predating Ian's Cochrane Review that I showed you in the BMJ publication in 2009, 201 00:25:33,450 --> 00:25:39,570 we published this review on Tamiflu on the basis of the swine flu epidemic. 202 00:25:40,410 --> 00:25:47,700 Remember when everybody was going to die? And actually, this is our first attempt at a rapid review. 203 00:25:49,180 --> 00:25:52,450 And we did this in about 6 to 8 weeks. 204 00:25:52,450 --> 00:25:57,669 And because of the emergence of the hotline and the need to produce evidence quickly, 205 00:25:57,670 --> 00:26:01,420 the BMJ really got its act together and published this within about four weeks. 206 00:26:01,810 --> 00:26:05,980 So from start to finish, we did this in about 12 weeks to journal publication. 207 00:26:06,730 --> 00:26:12,580 So in fact, rapid reviews are possible and we did one really early on. 208 00:26:12,610 --> 00:26:24,070 I actually have never articulated the story of why that happened, but it was cause it was a a policy priority also in my journey. 209 00:26:24,340 --> 00:26:31,300 We started to apply it in touched on this that you could apply systematic review methods to any type of questions if it's appropriate. 210 00:26:32,140 --> 00:26:35,230 And this was a visitor that had come to see me and Rafael. 211 00:26:35,230 --> 00:26:41,470 And we've looked at the adequacy of reporting, monitoring regimes in risk factors for cardiovascular disease. 212 00:26:42,010 --> 00:26:46,690 If you have hypertension, what do the guidelines say? And instead of looking at one guideline, 213 00:26:46,690 --> 00:26:55,990 we looked at 114 guidelines and couldn't find a single one that agreed in a monitoring regimen of 114 different global guidelines. 214 00:26:57,160 --> 00:27:05,560 And then finally, this one, as well as the idea that it was moving beyond just traditional treatment reviews, 215 00:27:05,770 --> 00:27:14,510 that you could look at different questions. And this is a very influential one, which was quite interesting in the guideline in the health guidelines. 216 00:27:14,530 --> 00:27:23,950 What it says is if you're assessing a child, it says under the age of two, he says a normal heart rate was above 120. 217 00:27:24,760 --> 00:27:29,560 And then the difference, if you were over to the difference was then 20 beats per minute different. 218 00:27:29,980 --> 00:27:35,590 So generally one day older, but the difference in your heart rate from normal to abnormal with 20 beats per minute. 219 00:27:35,590 --> 00:27:43,270 Well, that can't be right. You can't jump 20 beats per minute in one day and then you go from abnormal to normal. 220 00:27:44,500 --> 00:27:50,079 And as technology has emerged that we're better at measuring pulse, there was more of a need to go back and look at the evidence. 221 00:27:50,080 --> 00:27:54,010 And what they showed is we managed to produce with the systematic review evidence, 222 00:27:54,280 --> 00:28:01,930 smooth curve of heart rate and of respiratory rate in terms of child childhood infection. 223 00:28:03,160 --> 00:28:08,870 And so that's that was our second that's my second generation in the third generation. 224 00:28:08,890 --> 00:28:14,200 Moving on is the complexity of what we did changed and of what I did. 225 00:28:14,440 --> 00:28:17,860 Learning new skills meant you could move questions beyond. 226 00:28:17,860 --> 00:28:23,680 And this one is actually a systematic review of oral anticoagulation that we done the review for. 227 00:28:23,920 --> 00:28:31,240 But there was a need to do an individual patient data analysis, and that's a much more complex thing to do, 228 00:28:31,750 --> 00:28:37,150 much more costly, and you have to have a real specific need of why you want to do it. 229 00:28:38,140 --> 00:28:39,640 But it is very helpful. 230 00:28:39,880 --> 00:28:49,240 So we could answer all sorts of questions like we could do tying to event analysis, which weren't possible with just the published aggregated data. 231 00:28:49,630 --> 00:28:55,040 We could also look at number needed to treat year one, year to year three, year four in year five. 232 00:28:55,300 --> 00:29:00,180 And what we showed is the longer you had the intervention, the better you number needed to treat looked. 233 00:29:00,940 --> 00:29:02,830 So that was individual patient data. 234 00:29:03,520 --> 00:29:12,370 So in this increasing complexity, Ian pointed to this what we started to do use different ranges of techniques to answer questions. 235 00:29:12,790 --> 00:29:19,719 So this is qualitative studies here, a systematic review, a diagnostic review. 236 00:29:19,720 --> 00:29:24,940 I participated and here's another one looking at prognostic study. 237 00:29:25,120 --> 00:29:30,130 So the range of study type was expanding in terms of what you could do. 238 00:29:31,690 --> 00:29:39,340 And then finally, so this is what I call in my journey a fourth of first generation, second generation, third generation, 239 00:29:40,150 --> 00:29:47,050 fifth generation was about looking at randomised controlled trials and thinking about single technology, 240 00:29:47,230 --> 00:29:55,810 single intervention, comparing one with another, the second generation with actually expanding the ranges of study types. 241 00:29:56,350 --> 00:30:01,630 And then in the third generation, thinking about the complexity of the type of reviews, 242 00:30:02,800 --> 00:30:06,790 and then all the bets were off the table when we did this review. 243 00:30:07,980 --> 00:30:11,760 Because it took us five years to get the data. 244 00:30:12,330 --> 00:30:20,490 So in 2009, we published the Children's Review. I met Tom Jefferson for the first time in two nine because he was doing the adult review. 245 00:30:21,810 --> 00:30:28,440 And the reason we got together is the BMJ sent me a paper to review, which was done by Tom Jefferson. 246 00:30:29,280 --> 00:30:34,650 And it was the reanalysis of the Tamiflu data for the 2009 pandemic. 247 00:30:35,520 --> 00:30:39,180 And I said, Well, we know the results. But I said, Oh, my God. 248 00:30:39,870 --> 00:30:44,490 They found some incredibly important that eight of the ten trials were never published. 249 00:30:44,790 --> 00:30:50,129 This is the story. Forget the results. There's some really odd going on here and I didn't realise that. 250 00:30:50,130 --> 00:30:58,910 We didn't realise at the time we publish never even occurred to me that actually there was a huge issue of what we call under the waterline. 251 00:30:58,920 --> 00:31:02,850 There's all this evidence that exists that is completely missing. 252 00:31:03,810 --> 00:31:07,740 And that's been a very interesting journey. It's took is five years to get the data. 253 00:31:08,280 --> 00:31:13,830 And it changed my approach to how we can trust the evidence in publications. 254 00:31:17,060 --> 00:31:21,260 And then I'm just going to wind up with a couple of things that also made me think about the president. 255 00:31:21,290 --> 00:31:25,760 This was an article by Richard Smith, which was the Corporate Cocoon Collaboration at 20. 256 00:31:26,300 --> 00:31:29,840 It's published as a blog, but in their John Brodie we worked with. 257 00:31:30,410 --> 00:31:35,930 He was very forthright. People who've met John Brodie will find him opinionated, but he is a genius. 258 00:31:35,930 --> 00:31:40,009 He runs the trip database and in there they have a question and answering service 259 00:31:40,010 --> 00:31:43,700 for which there's collected thick files and questions from stakeholders, 260 00:31:43,700 --> 00:31:45,380 from professionals. There are questions. 261 00:31:46,010 --> 00:31:55,790 And in this article he says out of 358 questions asked in dermatology, only three could be answered by a single systematic review. 262 00:31:56,600 --> 00:32:03,079 So that makes you think about the complexity of the questions that we ask and with the evidence to date them questions. 263 00:32:03,080 --> 00:32:04,340 It's quite an interesting issue. 264 00:32:04,350 --> 00:32:15,290 So based on that, I'd been going back and forward with John quite a bit and asking him questions about how do we meet the agenda of Archie Cochrane? 265 00:32:15,290 --> 00:32:20,329 Because that's what he said to me. How do you meet this agenda? So there are some important issues. 266 00:32:20,330 --> 00:32:24,200 Periodically updating some people, including Cochrane, 267 00:32:24,410 --> 00:32:32,479 have decided that periodic updating is too difficult to achieve because as you grow the number of reviews, how do you keep them all today? 268 00:32:32,480 --> 00:32:35,510 And actually what they've gone for is a prioritisation strategy. 269 00:32:36,470 --> 00:32:41,240 What about the range of questions and different study types I've showed you? 270 00:32:41,570 --> 00:32:49,010 That's a problem also. And then we had in the clinical study reports, well, you can't do that or individual patient data for all reviews. 271 00:32:49,880 --> 00:32:52,280 So here's where we are in our thinking. 272 00:32:53,840 --> 00:33:04,790 Meeting the two agendas is that actually we could do a lot more in the Rapid Review arena, but use rapid reviews to update regular, 273 00:33:05,030 --> 00:33:12,860 respond to new emerging evidence support practice, support policy, keep existing reviews of today and automate the components. 274 00:33:14,480 --> 00:33:22,400 But I think there needs to be a middle bit, which I call the stakeholder prioritisation, where patients and people who are invested in the answer, 275 00:33:22,400 --> 00:33:32,240 like clinicians, prioritise whether you do a more thorough job because we just can't do all of the work on all of these reviews. 276 00:33:32,510 --> 00:33:39,230 So for instance, if the Rapid Review identifies there are two or three trials due to publish, why would you do the full systematic review? 277 00:33:40,640 --> 00:33:44,000 What happens if you find that the rapid review actually makes no difference? 278 00:33:44,330 --> 00:33:47,990 Why do all that work? If you answer this, will it make a difference to you? 279 00:33:49,550 --> 00:33:54,350 Why bother? In the robust review? They're much more labour intensive. 280 00:33:55,280 --> 00:34:01,260 They deal with much more data. There are 170,000 pages on our Cochrane Review in the in the CFR. 281 00:34:02,150 --> 00:34:05,000 It's going to take you two years to do and analyse that data. 282 00:34:06,020 --> 00:34:12,950 They should address publication and reporting bias, seek out important patient outcomes and analyse harms. 283 00:34:13,130 --> 00:34:20,630 Most of the problems that exist in the current evidence base that we seek to have problems such as cholesterol lowering are 284 00:34:20,630 --> 00:34:29,480 because of the problems with analysing harm and also they should seek to go beyond the globalisation using non-English studies. 285 00:34:29,480 --> 00:34:38,660 So they should not just rely on English language. And so I put that into the present is about four things. 286 00:34:39,170 --> 00:34:42,440 It's about how do we deal with reviews of single technologies. 287 00:34:43,760 --> 00:34:49,400 I think if you started on your systematic review journal where you start, if I would start with doing a review in that area, 288 00:34:49,670 --> 00:34:53,149 but then there's a movement to different study types which then go actually 289 00:34:53,150 --> 00:34:57,500 the range of reviews that required ones to study go beyond randomised trials, 290 00:34:58,100 --> 00:35:01,010 which differs from March's original definition. 291 00:35:02,030 --> 00:35:11,450 I do think we need to then have a process for integrating complex reviews, but that can't be for everything and that's where the rapid both reviews. 292 00:35:11,900 --> 00:35:17,809 And then the fourth is the sort of beyond journal approach, beyond the journal publication. 293 00:35:17,810 --> 00:35:22,190 And we see this increasingly that evidence is not in journals that we require. 294 00:35:22,970 --> 00:35:27,980 So before we move on to Kamaal, I want you to just share a story of what next? 295 00:35:30,260 --> 00:35:34,010 About a year ago, I got approached by a lady called Marie Lyon, 296 00:35:34,910 --> 00:35:40,610 who worked for the Association for the Childhood Health Damaged by Hormone Pregnancy Test. 297 00:35:40,610 --> 00:35:45,229 Can't remember the acronym, the Expert Working Group for the MHRA. 298 00:35:45,230 --> 00:35:49,460 It spent two years looking at hormone pregnancy test subjects called primitive, 299 00:35:51,260 --> 00:36:01,070 and they come to a conclusion that there was no there was no association or causation between hormone pregnancy tests and congenital malformations, 300 00:36:02,390 --> 00:36:07,280 and that they looked into the evidence and they come to the conclusion that it's very poor quality. 301 00:36:07,730 --> 00:36:12,980 And when they looked at it, there's no reason they can find any association. 302 00:36:13,820 --> 00:36:20,730 We got contacted by the patient group. Amory. He said, Could you look at this and tell us if the report is true? 303 00:36:21,540 --> 00:36:28,920 Then we will go away and put the case to rest. These treatments we used 40 years ago, they were a pregnancy test. 304 00:36:29,510 --> 00:36:34,620 Now you take two tablets of a pill of which one of the components contain progestogen. 305 00:36:34,620 --> 00:36:42,300 That was 40 times more stronger than the contraceptive pill you take today, ten times stronger than their morning after pill. 306 00:36:43,800 --> 00:36:48,240 It lasted for about 15 years and to me it's one of the stupidest ideas I've ever seen. 307 00:36:48,990 --> 00:36:53,640 And there are loads of issues in the BMJ and publications where they said this is a silly idea 308 00:36:53,880 --> 00:36:58,680 and we pick picking up a bit like Ian was showing lots of cases of congenital malformations. 309 00:36:59,370 --> 00:37:05,159 You may have picked this up in the perinatal some time in the literature and most of 310 00:37:05,160 --> 00:37:09,630 the arguments were about single studies that were either positive or not positive. 311 00:37:10,410 --> 00:37:17,070 Immediately I read the report. I said they failed to do one thing. 312 00:37:18,270 --> 00:37:23,190 Which I found incredibly odd. They failed to do a systematic review. 313 00:37:24,780 --> 00:37:28,950 A government organisation in 2017. 314 00:37:29,640 --> 00:37:34,450 Answering a question without doing a systematic review seemed autumn madness to me. 315 00:37:35,040 --> 00:37:39,240 So what did we do? We did the systematic review. We published a systematic review. 316 00:37:39,330 --> 00:37:44,480 Just been indexed on Pokemon. You can look at it, but interestingly, it doesn't make a difference for that. 317 00:37:44,490 --> 00:37:48,120 Suspending this email to to my collaborators. 318 00:37:48,300 --> 00:37:56,690 When you think about making a difference. The Department of Health and Social Care provided the following answer to your written 319 00:37:56,690 --> 00:38:01,280 parliamentary question to ask the Secretary of State for Health and Social Care if it will 320 00:38:01,280 --> 00:38:05,989 make an assessment of the metrics in the conclusions of the 2018 Report by Professor Karl 321 00:38:05,990 --> 00:38:10,850 Honeywood on oral hormone pregnancy tests and the risk of congenital malformations. 322 00:38:12,710 --> 00:38:19,580 Not only we ask the humans medicine is convening a new expert group to conduct an independent scientific review of our publication. 323 00:38:19,580 --> 00:38:24,200 So now that we conduct a review of our systematic review, and they're also, 324 00:38:24,200 --> 00:38:30,230 in addition asking the European Medicines Agency to actually conduct a review of our review. 325 00:38:31,670 --> 00:38:35,510 The reason I want to share that is because I think in this modern era, 326 00:38:35,540 --> 00:38:43,580 it's obvious that regulators should be doing systematic reviews on important questions like this are still failing to do that. 327 00:38:44,090 --> 00:38:50,360 So I think also in what next? There's not only that, there's a sort of need to really hit home. 328 00:38:50,360 --> 00:38:54,650 So it's interesting. I haven't actually seen the Academy of Medical Sciences policy paper. 329 00:38:55,880 --> 00:38:58,520 I think actually I'm going to put as well I didn't realise it, 330 00:38:58,520 --> 00:39:07,280 but I'm going to wonder what makes there's a real problem in regulatory science, not in our regulators, not using systematic reviews. 331 00:39:08,060 --> 00:39:16,160 So we are going to have to point that out, the absurdity of and so I think I'm going to finish at that point to say systematic 332 00:39:16,160 --> 00:39:20,540 reviews are just as important as they ever were and there's still a lot more work to do. 333 00:39:20,930 --> 00:39:48,060 Hand over to my colleague at that point, Monaco. Okay. 334 00:39:48,570 --> 00:39:52,770 So I'm going to chat for about 50 minutes or so in the future for this question. 335 00:39:52,920 --> 00:39:57,720 There's a lot I can cover, but in 15 minutes I just want to focus on three things. 336 00:39:58,020 --> 00:40:01,920 So as Carl said, I'm a commodity and I'm a GP. 337 00:40:01,950 --> 00:40:07,500 I work in East Oxford and I'm also director of an interim master's program in systematic reviews. 338 00:40:08,070 --> 00:40:14,330 So I may well be considered a bit biased about my love for this methodology, as would Ian and Carl. 339 00:40:14,880 --> 00:40:19,950 But don't take it from me. Let's go to the very top. So this is Chris Whitty. 340 00:40:20,190 --> 00:40:24,450 Chris Whitty is the UK chief scientific adviser for the Department of Health and Social Care. 341 00:40:25,650 --> 00:40:29,610 He's also head of and I Chop. And a few years ago Chris Whitty published this paper. 342 00:40:30,390 --> 00:40:34,140 What makes an academic paper useful to influence health policy? 343 00:40:36,000 --> 00:40:44,070 And this is what he wrote way ahead of any other academic contribution to policymaking is rigorous and unbiased synthesis of current knowledge. 344 00:40:45,780 --> 00:40:49,769 And he sort of threw down the gauntlet to everyone to see if the academic community 345 00:40:49,770 --> 00:40:53,850 as a whole could do one thing to improve the pathway from research to policy. 346 00:40:54,240 --> 00:40:58,560 It would be to improve the status, quality and availability of good synthesis. 347 00:40:58,980 --> 00:41:02,070 So that's coming from the chief scientific adviser in the UK. 348 00:41:02,550 --> 00:41:07,110 So it's not just us who's interested in this topic. Chris Whitty and his team as well. 349 00:41:09,900 --> 00:41:14,910 So let me talk about three things that I feel around growing the field of evidence census. 350 00:41:14,970 --> 00:41:20,670 I want to talk about embracing new methods and techniques, creating training opportunities and building capacity, 351 00:41:20,670 --> 00:41:25,290 including what we're doing here in the CBN to start with new methods and techniques. 352 00:41:26,280 --> 00:41:30,150 So why do we need new methods? Why do we need new techniques? There's a whole bunch of reasons. 353 00:41:30,330 --> 00:41:36,690 I'm going to give you three and some of them have been alluded to already. So, number one, we've got this huge problem with publication bias. 354 00:41:37,590 --> 00:41:41,280 I call talked about it already with regards to the Tamiflu review. 355 00:41:41,730 --> 00:41:46,800 We know that a large proportion, estimated 50% of studies clinical trials are not published. 356 00:41:47,580 --> 00:41:51,040 You're then going to do a systematic review of published studies. 357 00:41:51,060 --> 00:41:56,280 How confident do you feel about the estimates that you come to knowing that a large proportion of data is missing? 358 00:41:57,210 --> 00:41:59,730 So we've got this huge public problem about publication bias, 359 00:41:59,970 --> 00:42:07,710 and it takes techniques like Carl alluded to with the Tamiflu review of spending years trying to get this data, these clinical study reports. 360 00:42:07,980 --> 00:42:11,040 And we're doing a number of other reviews using this sort of data. 361 00:42:16,270 --> 00:42:23,739 Real-World comparisons often. One of the challenges is that we usually trials where they're comparing a 362 00:42:23,740 --> 00:42:29,830 treatment to a placebo or they're using outcomes which are surrogate outcomes. 363 00:42:30,400 --> 00:42:36,610 And actually, what you really want to know in the real world is, is this treatment better than what this patient is currently being taken? 364 00:42:37,790 --> 00:42:42,080 And also, is the outcome something that really matters to the patient rather than something that's a surrogate? 365 00:42:42,740 --> 00:42:47,090 So we need better ways to synthesise data to try and get that sort of data out of it. 366 00:42:49,730 --> 00:42:53,900 But find out on this list of three things complex, critical and policy. 367 00:42:53,900 --> 00:43:00,559 So we no clinical questions. Clinical policy are getting more complex and therefore we need better techniques. 368 00:43:00,560 --> 00:43:03,770 We need more advanced techniques to ask more than just does it work, 369 00:43:04,190 --> 00:43:09,530 but actually does it work in the context and does it work in the contextualisation of other problems? 370 00:43:09,530 --> 00:43:12,890 We've gotten issues like decision making. How do we make a decision with patients? 371 00:43:13,680 --> 00:43:20,780 I'll give you an example. So here Will is a common headline from the BBC Health about our problem with obesity. 372 00:43:21,140 --> 00:43:26,200 UK the most overweight country in Western. A Cochrane review. 373 00:43:26,830 --> 00:43:33,580 This one, which is a pretty good review, will tell us that actually exercise and good diet will help you lose weight. 374 00:43:35,200 --> 00:43:38,890 Right. And no one's update to that review because they're pretty confident about those results. 375 00:43:40,630 --> 00:43:44,470 Let's take that into context. Now, this is deliberately small for a reason, as you'll see in a second. 376 00:43:44,830 --> 00:43:49,670 So this is this is going to be a map of what contributes to the factors around obesity. 377 00:43:49,690 --> 00:43:55,750 It's a very complex issue and came out with a nice diet and physical activity. 378 00:43:56,440 --> 00:43:59,020 So those are the the components that we picked up in that review. 379 00:44:00,100 --> 00:44:04,540 But actually look at the other factors that are thought to contribute to the problem of obesity. 380 00:44:05,920 --> 00:44:09,460 So this is something from the full site report carried out in the mid 2000, 381 00:44:09,700 --> 00:44:18,520 which mapped all the other issues around obesity ranging from physiology, physical activity, environment, as you see, social psychology and so on. 382 00:44:18,970 --> 00:44:24,970 So we have to be able to synthesise a lot of this data and contextualise it to answer a very complex question. 383 00:44:24,970 --> 00:44:32,260 You know, asking, you know, what? How are we going to, you know, improve the management of obesity is a complex question, as you can see. 384 00:44:32,710 --> 00:44:38,080 So we need better techniques to answer questions, complex issues like obesity, for example. 385 00:44:39,250 --> 00:44:44,810 Okay. So as I'm alluding to, we're pretty good at being able to answer questions around, does something work? 386 00:44:44,830 --> 00:44:49,090 We've got good techniques. They've been honed mostly over a number of years. 387 00:44:49,300 --> 00:44:56,140 So pretty good to be able to answer that question. But what's happened is, as Ben alluded to already, we've got an emerging range of techniques now. 388 00:44:57,280 --> 00:45:02,620 So we now can explore questions around how accurate is a test we can explore and we can do 389 00:45:02,620 --> 00:45:07,029 scoping reviews to explore what sort of evidence is out there before we do a more formal, 390 00:45:07,030 --> 00:45:12,880 perhaps a more robust review. We've got things like really sciences which will use them will theory based approach 391 00:45:13,120 --> 00:45:17,830 to synthesising data and often from other sources that are published domains, 392 00:45:18,580 --> 00:45:22,510 qualitative senses, for example, to explore experience of using a treatment. 393 00:45:22,930 --> 00:45:24,940 And we've got rapid or restrictive reviews as well, 394 00:45:25,090 --> 00:45:30,740 whereby if you need a data or an evidence synthesis more rapidly, for example, you can use those methods. 395 00:45:31,210 --> 00:45:37,750 And we've got things like network match analysis whereby you can infer indirect comparisons between treatment A, 396 00:45:37,850 --> 00:45:41,020 treatment B when they weren't otherwise compared to each other in the original trials. 397 00:45:41,470 --> 00:45:46,660 So there's a whole host of new methods emerging. And if you are really, really interested in this in this work, 398 00:45:46,660 --> 00:45:53,260 I would recommend having a look at this paper published a couple of years ago by Andrea Tirico and her team in Canada. 399 00:45:54,100 --> 00:45:59,680 And what they do is they scope the literature to say what sort of new techniques in evidence synthesis are emerging. 400 00:46:00,760 --> 00:46:07,870 And they concluded they found at least 25 unique emerging methods in the evidence in this field, and probably by this time, there's going to be more. 401 00:46:08,410 --> 00:46:14,379 So this is an evolving field, and it's something that is worth paying attention to because as I said, 402 00:46:14,380 --> 00:46:23,080 it's changing as a result of a number of other factors. Carl alluded to in his final just fine a slide about the fact that we need complex reviews. 403 00:46:24,130 --> 00:46:28,810 This is something that is of interest to us because, you know, as I said, the data is getting more complex. 404 00:46:28,810 --> 00:46:30,190 We need more sophisticated methods. 405 00:46:30,550 --> 00:46:34,960 But actually, what we found when we looked into it, no one's actually identified or described what a complex review is. 406 00:46:35,740 --> 00:46:44,320 So what we did last year was we had a go, so we published a paper in the BMJ, UBL, and we defined the complex review as such. 407 00:46:44,530 --> 00:46:50,200 I'll just read it. So a systematic review performed by a multi-disciplinary team consisting of multiple components, 408 00:46:50,530 --> 00:46:53,890 large amounts of data from different sources or different perspectives, 409 00:46:54,280 --> 00:46:58,630 collectively contributing more than would be expected from the individual contributions, 410 00:46:59,080 --> 00:47:03,370 the individual components not being easily coordinated, analysed or disentangled. 411 00:47:04,450 --> 00:47:09,610 It's a starting point, but we couldn't believe that. No one's actually just describe what a complex review is. 412 00:47:10,000 --> 00:47:18,780 So we thought we'll have a code and we'll start the debate. Okay. 413 00:47:19,440 --> 00:47:23,339 So number two, the need to try to create more training opportunities. 414 00:47:23,340 --> 00:47:26,850 I said I direct the masters and you can see all three from inside. 415 00:47:26,850 --> 00:47:33,899 I'm a bit biased. I believe everyone, all health researchers, as Ian's alluding to an issue leading to all researchers, 416 00:47:33,900 --> 00:47:38,130 should consider doing this, this factory as part of the first part of their training. 417 00:47:39,570 --> 00:47:45,719 But it's not just me who thinks that this is the research arm of the NHS, the nature, the fund. 418 00:47:45,720 --> 00:47:50,580 A lot of our research every year they put out a fellowship call for systematic 419 00:47:50,580 --> 00:47:55,020 review training and these would be fellowships specifically for these reasons, 420 00:47:55,380 --> 00:47:58,320 to address the current shortage of systematic reviews in the UK. 421 00:47:58,950 --> 00:48:04,380 So we're talking about Chris Whitty saying this is the most important study designed to change policy, 422 00:48:04,650 --> 00:48:10,080 but we don't have enough systematic reviews in the UK. So now I'll put out these sorts of fellowships. 423 00:48:12,790 --> 00:48:14,920 We also, as many of you or some of you may know, 424 00:48:15,100 --> 00:48:21,249 we also run a master's in evidence based health care and a natural offshoot from that a couple of years ago was a 425 00:48:21,250 --> 00:48:28,330 dedicated MSC in systematic reviews and as part of one of the multiple people is specifically on complex reviews. 426 00:48:29,090 --> 00:48:33,760 And as a as I said, as we say here, developing a better understanding of the broader forms of evidence. 427 00:48:34,330 --> 00:48:37,330 And we had our first cohort of students on there last year. 428 00:48:39,340 --> 00:48:44,500 We do one day workshops as well. The CPM got one coming up in March and I was interested as well. 429 00:48:46,970 --> 00:48:47,480 And then finally, 430 00:48:47,480 --> 00:48:54,020 I just thought and by telling you a little bit more about some of the work we're doing to build capacity beyond some of the things I've said already. 431 00:48:56,600 --> 00:49:01,340 So if you haven't visited our website, you'll see. Please do CBN dot net. 432 00:49:01,700 --> 00:49:08,300 You'll see that we do a lot of systematic reviews and in fact, we promise this is an exit from our research strategy. 433 00:49:10,160 --> 00:49:13,900 So from that point of. Now. 434 00:49:15,120 --> 00:49:21,270 Okay, so CPM research is a global source of high quality evidence on which clinical decisions can be readily based. 435 00:49:22,260 --> 00:49:23,339 And we prioritise. 436 00:49:23,340 --> 00:49:29,900 If you look here, the first thing we raise, we will generate and synthesise high quality evidence that benefits patients and society. 437 00:49:29,960 --> 00:49:32,910 That's that is part of our primary, part of our vision. 438 00:49:32,910 --> 00:49:39,900 And we do that across a number of areas prevention, diagnostics, communicating evidence methods, therapeutics and health services. 439 00:49:40,560 --> 00:49:44,160 And as I said, we produce a lot of systematic reviews. 440 00:49:44,970 --> 00:49:52,380 So a natural offshoot is that is I'm pleased to announce that we will now have a dedicated EPM synthesis unit as part of the CBN. 441 00:49:52,890 --> 00:49:59,460 And here's some of our ongoing reviews. So we've got David Noonan leading a review on the role of physical activity in chronic disease. 442 00:49:59,760 --> 00:50:02,970 We've got some we've got a review on the role of paramedics in primary care. 443 00:50:03,690 --> 00:50:10,410 But as part of also doing reviews, we're also developing and pushing forward the tools available for others as well. 444 00:50:11,220 --> 00:50:16,290 And one of the things that we've been working on is a toolkit and evidence of this toolkit, 445 00:50:16,590 --> 00:50:22,139 because, as I said, with this growing number of review designs and methods and as I said, 446 00:50:22,140 --> 00:50:25,650 there's more coming, how are you going to keep abreast with the how are we going to how are we going to be 447 00:50:25,650 --> 00:50:30,480 able to know which which type of synthesis design will be for which sort of question? 448 00:50:31,470 --> 00:50:42,670 So that's what we're doing. We're developing a toolkit. And we are describing each review type, describing what it's for, 449 00:50:42,730 --> 00:50:48,340 how we use the mathematical framework and how you can apply it in your synthesis design. 450 00:50:52,560 --> 00:51:00,150 So that's just a short summary of what we're doing around new methods, training opportunities and building capacity as well. 451 00:51:00,270 --> 00:51:06,720 And I'll just leave you with this final quote from Ben Goldacre, who also works closely with us at the centre of the Space Medicine. 452 00:51:07,140 --> 00:51:08,430 The notion of the systematic review, 453 00:51:08,430 --> 00:51:14,430 looking at the totality of evidence is currently one of the most important innovations in medicine over the past 30 years. 454 00:51:16,070 --> 00:51:16,430 Thank you.